Every fact, on the record.

“In the UK, the National Living Wage rose to £12.71 an hour in April 2026 for workers aged 21 and over.”

Confirmed: the NLW rose by 4.1% to £12.71 on 1 April 2026.

The NLW is the statutory minimum for workers aged 21+, set on Low Pay Commission advice, with two-thirds of median hourly earnings as the reference point. In April 2026 the 18-20 rate became £10.85 and the 16-17 and apprentice rates £8.00.

“The NLW rises by 4.1 per cent in April 2026 to £12.71.”

“When the UK brought in the minimum wage in 1999, critics said it would destroy jobs, but employment went on to hit record highs.”

Confirmed: the Low Pay Commission's 20-year review found no strong evidence of job losses from the minimum wage, and employment rates reached record levels.

The minimum wage started on 1 April 1999 at £3.60 an hour for workers aged 22+. By 2019 the LPC had commissioned more than 30 studies on jobs and hours. Some found small falls in hours, but the results were not consistent. Record employment does not by itself prove the minimum wage had no effect.

“rather than destroy jobs, as was originally predicted, we now have record employment rates”

“The UK's Low Pay Commission says recent National Living Wage increases have not significantly harmed employment.”

Confirmed: the LPC's 2025 report judged that recent NLW increases had no significant negative impact on employment.

The same report says that between 2016 and 2025 the real weekly pay of NLW earners rose 20%, against 4% for the median job, and that the NLW's bite reached a record high. It also warns that the April 2025 rise coincided with higher employer National Insurance, which makes effects harder to separate.

“Considering all the evidence, our judgement is that recent NLW increases have not had a significant negative impact on employment.”

“IFS research found the National Living Wage's introduction and early rises had no statistically significant effect on employment.”

Confirmed: for the 2016-2019 upratings, the IFS central estimate was a 0.1% fall in employment per uprating, with a 95% confidence interval of -0.4% to +0.2%. This was not statistically significant.

The study compared low-wage and high-wage regions of the UK. It found large wage gains and spillovers up to about the 20th percentile. It also found some borderline-significant evidence of a larger negative effect on women's employment, of about -0.44%.

“any impacts on employment of the introduction of the NLW and its upratings were small, and not statistically significantly different from zero”

“A major US study of 138 state minimum wage rises found that the number of low-wage jobs was essentially unchanged five years later.”

Confirmed: Cengiz, Dube, Lindner and Zipperer studied 138 state-level increases between 1979 and 2016 and found no change in the overall number of low-wage jobs over five years.

The study is a bunching analysis published in the Quarterly Journal of Economics (2019). Jobs paying below the new minimum disappeared and were matched by new jobs at or just above it. The authors did find some reduced employment in tradable sectors.

“We find that the overall number of low-wage jobs remained essentially unchanged over five years following the increase.”

“In the UK, the share of low-paid jobs has fallen to its lowest level on record.”

Confirmed: 2.5% of employee jobs were low-paid in April 2025, down from 3.4% in 2024 and the lowest since the series began in 1997.

The ONS defines low pay as below two-thirds of median hourly earnings, which was £11.97 in April 2025. The NLW for workers aged 21 and over (£12.21 then) was above that threshold. Low pay was most concentrated among employees aged 16 to 21, part-time jobs and hospitality. The figures cover jobs paid at adult rates, which excludes trainee and junior rates, and the 2025 estimates are provisional.

“The proportion of low-paid employee jobs fell to 2.5% in April 2025 from 3.4% in 2024, the lowest since the series began in 1997”

“The OBR forecast that the National Living Wage would cost around 60,000 jobs.”

Confirmed: in November 2015 the OBR estimated that unemployment would be around 60,000 higher than without the NLW.

This was a forecast built on an assumed response of hours worked to higher labour costs, split between fewer jobs and shorter hours. It was not a measured outcome. Later LPC and IFS evaluations found no statistically significant employment loss from the NLW.

“unemployment would be around 60,000 higher than in the absence of the NLW”

“The OBR estimated that raising the National Living Wage to two-thirds of median earnings would push unemployment up by about 140,000.”

The OBR gave this figure in October 2018, but when the two-thirds target was adopted in March 2020 it cut its estimate to a rise in unemployment of around 50,000 by 2024.

In October 2018 the OBR illustrated that a two-thirds NLW could raise unemployment by 0.4 percentage points, about 140,000 people, plus a similar cut in hours. In March 2020, citing limited evidence that past rises had significantly affected employment, it lowered its assumed responsiveness of hours to pay. Both were model estimates, not measured outcomes. The LPC has not found a significant negative employment impact from recent NLW rises.

“That corresponds to a rise in unemployment of around 50,000”

“When Seattle raised its minimum wage to $13, low-wage workers lost hours and ended up with less pay overall.”

Contested: a University of Washington study (2017, revised 2018) found the rise to $13 in 2016 cut hours in low-wage jobs by 6-7% and pay per low-wage job by $74 a month, but the same team's 2018 follow-up found workers already in low-wage jobs gained $8-12 a week on average.

Seattle raised its minimum from $9.47 to up to $11 in 2015 and up to $13 in 2016. In the follow-up, all of the earnings gains went to more experienced workers; less-experienced workers saw no significant change in weekly pay, and fewer new workers entered the workforce. The team's 2022 journal article says its aggregate analysis likely overstates employment effects. The findings are specific to one city.

“On net, the minimum wage increase from $9.47 to as much as $13 per hour raised earnings by an average of $8-$12 per week.”

“The US Congressional Budget Office estimated that a $15 federal minimum wage would cost about 1.4 million jobs.”

Confirmed as CBO's average estimate for 2025 in its February 2021 analysis. The same analysis estimated that 900,000 fewer people would be in poverty, and 17 million workers would be directly affected.

CBO put a 33% chance on job losses between zero and 1 million, and a 33% chance on losses between 1 million and 2.7 million. The US federal minimum has been $7.25 since July 2009.

“the number of people in poverty would decline by 900,000, but employment would also drop by 1.4 million people, according to CBO's average estimate”

“Youth employment in the UK is falling, with about 69,000 fewer employees under 25 in the year to August 2025.”

The 69,000 fall was for the year to August 2025. ONS early estimates show a much smaller fall of 15,000 payrolled employees under 25 in the year to August 2026.

The Low Pay Commission reported that employees aged under 25 fell by around 69,000 (1.8%) in the year to August 2025. The latest figure is an early estimate that may be revised; total payrolled employees fell by 145,000 (0.5%) over the same year. Young people are concentrated in hospitality and retail, and in April 2026 the LPC said it was not aware of any robust evidence that higher youth rates had contributed to the falls.

“Between August 2025 and August 2026, there was a decrease of 15,000 payrolled employees aged under 25 years.”

“Most minimum wage workers in the US are teenagers earning pocket money.”

In 2024, workers under 25 (not just teenagers) made up 43% of those paid at or below the US federal minimum. A majority were 25 or older.

Only about 1.0% of US hourly workers earned $7.25 or less in 2024: 82,000 at exactly $7.25 and 760,000 below it, some of whom are exempt from minimum wage laws. About two-thirds worked in leisure and hospitality. Many states set higher minimums.

“they accounted for 43 percent of those paid the federal minimum wage or less”

“The CBO says a $15 minimum wage would put close to 4 million Americans out of work.”

CBO's 2019 median estimate was 1.3 million jobs lost. The 3.7 million figure was the top of a range, with a two-thirds chance of losses between about zero and 3.7 million.

Senator Rand Paul made this claim in 2021, citing the upper bound as if it were the forecast. In the same analysis CBO estimated that 1.3 million people would move out of poverty and 17 million workers would get higher pay.

“refers to the high end of CBO's range of potential outcomes”

“Raising the minimum wage drives up inflation across the whole economy.”

The UK Low Pay Commission finds the minimum wage is a major cost driver in some industries, but its effect on headline inflation is negligible.

Firms in low-paying sectors often pass minimum wage rises on through higher prices, and the LPC says this has become a more common response. So price rises in those sectors are real, but the LPC finds the NLW has only a minimal effect on overall inflation.

“while the minimum wage is an important cost driver in some industries, its impact on headline measures of inflation is negligible”

“Raising the minimum wage is the best-targeted way to help the poorest households in the UK.”

The IFS found that the biggest cash gains from NLW rises went to middle-income households. The bottom and top 10% each gained under 35p a week per uprating, and in the 2nd and 3rd income deciles almost half of the earnings gain was clawed back through reduced benefits and higher taxes.

Some of the lowest-income households have no one in paid work, so they gain nothing directly. Looking only at households with someone in work, the gains are more progressive.

“the biggest cash gains go to the middle of the household income distribution”

“The famous Card and Krueger study proved that New Jersey's minimum wage rise actually increased fast-food jobs.”

Card and Krueger's study (1993 working paper, published 1994) found New Jersey fast-food employment rose 13% relative to Pennsylvania after the minimum went from $4.25 to $5.05 in April 1992. A re-analysis of payroll records from 230 restaurants by Neumark and Wascher found a 4.6% relative fall instead.

Card and Krueger's own 1998 reanalysis of BLS ES-202 payroll data found similar or slightly faster job growth in New Jersey than in eastern Pennsylvania, and attributed the Neumark-Wascher difference to a small set of restaurants owned by one franchisee. It remains a single, contested case study that supports small or no job losses from this rise more than it proves that minimum wage rises create jobs.

“the New Jersey minimum wage increase led to a 4.6 percent decrease in employment”

“In Canada, more than 70% of the cannabis people consume now comes from legal sources, up from about a fifth when legalisation began in 2018.”

Confirmed: legal (licensed medical and non-medical) sources now supply a little over 70% of the value of cannabis consumed in Canada, up from 22% in the fourth quarter of 2018, when legalisation began.

Statistics Canada measures this by the value of cannabis consumed (household spending), not by the number of users. In October 2023 it reported that the legal share had passed 70% by the first half of 2023. Its revised quarterly spending data now put the legal share at about 69% for 2023 and about 72% for 2025.

“more than 70% of the total value of cannabis consumed in Canada was from a legal (licensed medical and non-medical) source”

“Nearly nine in ten Americans say cannabis should be legal for medical or recreational use, and 24 US states have legalised recreational use.”

Confirmed: in a January 2026 Pew survey, 55% backed legal medical and recreational use and 33% medical use only, while 11% said it should not be legal at all. 24 states plus DC allow recreational use.

Pew says 53% of Americans live somewhere recreational cannabis is legal. Recreational cannabis is still illegal under US federal law, although in April 2026 the US Justice Department moved FDA-approved and state-licensed medical marijuana from Schedule I to Schedule III.

“Just 11% say the drug should not be legal at all.”

“Germany legalised recreational cannabis for adults in April 2024, allowing people to possess up to 25 grams.”

Confirmed: Germany's Cannabis Act came into force on 1 April 2024. Adults may legally carry up to 25 g, and may keep up to 50 g at home.

Each adult may also grow up to three plants at home for personal use, and non-commercial 'cannabis clubs' of up to 500 members may grow cannabis and share it with members. It is a non-commercial model, not a North American-style retail market.

“Possession of amounts up to 25 g is legal for adults.”

“Early evidence from Germany found no significant rise in cannabis use or drug-driving in the first eight months after legalisation.”

Confirmed: a study comparing Germany with Austria found no significant short-term effect on cannabis use or driving under the influence of cannabis eight months after legalisation.

Past-year use in Germany rose from 12.1% to 14.4%, but that trend was not significantly different from Austria's. This is short-term evidence only; the authors call for longer monitoring of drug-driving and traffic data.

“Eight months after legalisation, no significant short-term effects on cannabis use or DUIC were observed.”

“Even under prohibition, about 1 in 15 adults aged 16 to 59 in England and Wales used cannabis in the past year.”

Confirmed: in the year ending March 2025, 6.5% of people aged 16 to 59 said they had used cannabis in the last 12 months, including 12.5% of 16 to 24 year olds.

These figures come from the Crime Survey for England and Wales. Cannabis has been the most used illegal drug in England and Wales every year since estimates began in 1995. Use among 16 to 24 year olds is lower than it was in 2015.

“6.5% of people aged 16 to 59 years and 12.5% of people aged 16 to 24 years reported having used the drug”

“In the UK, cannabis is a Class B drug and possession can carry up to five years in prison.”

Confirmed: cannabis is Class B. The maximum penalty is 5 years for possession and 14 years for supply or production, with an unlimited fine possible in both cases.

These are maximum sentences. The same government page says police can give a warning or an on-the-spot fine of up to £90 to someone found with cannabis.

“Up to 5 years in prison, an unlimited fine or both”

“Daily use of high-potency cannabis is linked to nearly five times the odds of psychosis, and about 30% of new psychosis cases in London could be prevented if high-potency cannabis were not available.”

Confirmed: the EU-GEI study found an adjusted odds ratio of 4.8 for daily high-potency use, and estimated that 30.3% of first-episode psychosis cases in London (50.3% in Amsterdam) could be prevented.

This was a case-control study of 901 first-episode psychosis patients and 1,237 controls at 11 sites, published in Lancet Psychiatry in 2019. High potency meant 10% THC or more. The 'preventable' figures assume that the link is causal.

“rising to 30·3% (15·2-40·0) in London and 50·3% (27·4-66·0) in Amsterdam”

“A Danish study estimated that up to 30% of schizophrenia cases in men aged 21 to 30 could have been prevented by avoiding cannabis use disorder.”

Confirmed: using Danish health records, researchers estimated that as many as 30% of schizophrenia cases among men aged 21 to 30 might have been prevented by averting cannabis use disorder.

The study, published in Psychological Medicine in 2023 and co-led by NIDA, covered more than 6 million people across five decades. It concerns cannabis use disorder (problematic, dependent use), not all cannabis use. The estimate comes from statistical modelling of an association.

“as many as 30% of cases of schizophrenia among men aged 21-30 might have been prevented by averting cannabis use disorder”

“In Ontario, Canada, the share of new schizophrenia cases linked to cannabis use disorder almost tripled as cannabis was liberalised and then legalised.”

Confirmed: the share of schizophrenia cases associated with cannabis use disorder rose from 3.7% before legalisation to 10.3% in the legalisation period.

This 2025 cohort study covered 13.6 million Ontarians aged 14 to 65. Overall schizophrenia incidence stayed stable, and the rise was steady rather than jumping at each policy change. It shows an association, not proof that legalisation caused the rise.

“the proportion of incident cases of schizophrenia associated with CUD almost tripled during a period of substantial liberalization of cannabis policy”

“Commercial legal cannabis markets in Canada and the US have been linked to more adult cannabis use, more cannabis use disorder and stronger products.”

Confirmed: a 2026 review of international cannabis policy changes from 2000 to 2025 found that commercial legal markets were associated with higher adult use, more cannabis use disorder and rising potency.

The same review found weaker links for tightly regulated models that limit commercialisation, and little evidence of harm from decriminalisation. Critics of commercial markets and supporters of strict regulation can both cite it.

“Commercialised legal markets for non-medical use in Canada and the USA were associated with increased prevalence of cannabis use and cannabis use disorder in adults”

“More Americans now use cannabis daily or near-daily than drink alcohol that often: 17.7 million versus 14.7 million in 2022.”

Confirmed: the 2022 US national drug use survey recorded 17.7 million daily or near-daily cannabis users and 14.7 million daily or near-daily drinkers, the first time cannabis came out ahead.

Far more people drink overall, but heavy, frequent use is much more common among cannabis users. From 1992 to 2022, the per-person rate of daily or near-daily cannabis use rose 15-fold. The author links these trends to policy liberalisation.

“the 2022 survey, for the first time, recorded more daily and near daily users of cannabis than alcohol (17.7 vs. 14.7 M)”

“About 3 in 10 people who use cannabis have cannabis use disorder.”

Confirmed: the US CDC estimates that approximately 3 in 10 people who use cannabis have cannabis use disorder.

The CDC says the risk is higher for people who start young and use often. It also says average THC concentration almost doubled, from 9% in 2008 to 17% in 2017. This is a US estimate.

“Approximately 3 in 10 people who use cannabis have cannabis use disorder.”

“Legalising medical cannabis reduces opioid overdose deaths.”

An early study linked US state medical cannabis laws to lower opioid overdose deaths from 1999 to 2010. Repeating the same analysis through 2017 flipped the association from -21% to +23%, and the authors concluded the link was probably spurious.

This comes from a 2019 PNAS study that re-ran the original 2014 analysis. It also found no link between recreational laws and opioid deaths. The authors say claims that cannabis laws cut overdose deaths 'should be met with skepticism'.

“the association between state medical cannabis laws and opioid overdose mortality reversed direction from −21% to +23%”

“Legalising recreational cannabis leads to more teenagers using it.”

The evidence is mixed. A 2024 meta-analysis of US studies linked recreational legalisation to modestly higher odds of past-month cannabis use among young people, more so among young adults than adolescents. But a 47-state study of nearly 900,000 high school students found no net increase through 2021.

In the meta-analysis's more recent studies, the odds of past-month use were about 9% higher for adolescents and 22% higher for young adults. The high school study (JAMA Pediatrics, 2024) found retail sales were linked to more teens not using at all, but to more frequent use among teens who did use. A 2026 Northern California study of 1.3 million teen screening questionnaires found use rose after legalisation, then fell after the COVID-19 pandemic, returning to pre-legalisation levels by 2024.

“Our meta-analytic results showed modest positive effects of RCL on past-month cannabis use (more so in young adults than in adolescents)”

“Legalisation has caused a surge in schizophrenia and other psychotic disorders.”

A 2026 review of cannabis policy changes worldwide found no consistent evidence that they changed how common psychotic disorders are. It did find that commercial legalisation was linked to more hospital admissions for psychosis, including psychosis alongside cannabis use disorder.

Individual-level studies link daily use of high-potency cannabis to psychosis (see CB-07). In Ontario, schizophrenia incidence stayed stable, but new diagnoses of unspecified psychosis were about 84% higher in 2018-2022 than in 2006-2015, and the share of schizophrenia cases linked to cannabis use disorder rose (see CB-09). These are associations, not proof that legalisation caused the changes.

“There was no consistent evidence for associations between policy change and the prevalence or incidence of psychotic disorders.”

“Legalisation wipes out the illegal cannabis market.”

Legalisation shrank Canada's illegal market sharply but did not eliminate it. Legal sources went from 22% of the value of cannabis consumed in late 2018 to about 70% within five years, leaving roughly 30% to unlicensed sellers.

Statistics Canada's revised quarterly spending data show unlicensed sources still took about 28% of household cannabis spending in 2025, roughly $2.6 billion. These are Canadian figures; results in other countries may differ.

“more than 70% of the total value of cannabis consumed in Canada was from a legal (licensed medical and non-medical) source”

“Nuclear power is one of the safest energy sources, causing far fewer deaths per unit of electricity than fossil fuels.”

Confirmed: Our World in Data estimates about 0.03 deaths per terawatt-hour for nuclear (including Chernobyl and Fukushima), against 24.6 for coal and 2.8 for gas, making nuclear 99.8% safer than coal and 97.6% safer than gas.

The death rates count air pollution and accidents across the supply chain. Wind (about 0.04) and solar (about 0.02) are similarly safe; the differences between nuclear, wind and solar are within the uncertainty.

“Nuclear energy, for example, results in 99.9% fewer deaths than brown coal; 99.8% fewer than coal; 99.7% fewer than oil; and 97.6% fewer than gas.”

“Over its full lifecycle, nuclear power emits about as little carbon as wind and far less than gas.”

Confirmed: the IPCC puts median lifecycle emissions at 12 g CO2-equivalent per kWh for nuclear and 11 g for onshore wind, against 490 g for gas and 820 g for coal.

Lifecycle figures cover mining, fuel processing, construction and decommissioning, not just operation. Estimates vary by study: the IPCC range for nuclear runs from 3.7 to 110 g, and Our World in Data, using newer UN (UNECE) figures, finds coal emits more than a hundred times more than nuclear.

“Oil and gas are also much worse than nuclear and renewables but to a lesser extent than coal.”

“In the US, nuclear plants run at full power more than 90% of the time, far more than wind or solar.”

Confirmed: the US Energy Information Administration says the average capacity factor of US nuclear plants was 91% in 2025, higher than for any other type of power plant; wind was about 34% and solar photovoltaic about 24%.

Capacity factor is actual output as a share of potential output. Wind and solar generate only when the wind blows or the sun shines, while gas and coal plants are used flexibly to meet demand (about 58% for gas combined-cycle and 49% for coal in 2025). The 2025 figures are preliminary; EIA's final figure for 2024 was 90.8%, while the Department of Energy's nuclear office cites more than 92% for 2024. Nuclear supplied about 17% of US electricity generation in 2025.

“The average annual capacity factor for nuclear power plants in 2025 was 91%”

“The UK government has set a goal of up to 24 GW of nuclear power by 2050, enough for about a quarter of the country's electricity.”

Confirmed: the January 2024 Civil Nuclear Roadmap set out an ambition to increase nuclear generation up to fourfold, to 24 GW by 2050, described as enough for a quarter of UK electricity needs. The same 24 GW ambition had already been set out in the government's April 2022 Energy Security Strategy.

Both announcements came from Conservative governments. The Labour government's November 2025 response on nuclear planning policy (EN-7) says energy National Policy Statements set no limit or target for low-carbon infrastructure, and points instead to funded commitments: £14.2 billion for Sizewell C and over £2.5 billion for a first small modular reactor programme. The World Nuclear Association reported in January 2026 that the government had not indicated any change to the 24 GW goal. Hinkley Point C is also under construction.

“increase generation of this homegrown supply of clean, reliable, and abundant energy by up to 4 times to 24 gigawatts (GW) by 2050”

“All the spent fuel US nuclear plants have ever produced would fit on a single football field less than 10 yards deep.”

Confirmed: the US Department of Energy said in 2022 that the roughly 90,000 metric tons of spent fuel produced by US commercial reactors since the 1950s could fit on one football field at a depth of under 10 yards.

The US generates about 2,000 metric tons of spent fuel a year, and a July 2025 Department of Energy presentation put the total at about 95,000 metric tons. The small volume is real, but the US still has no permanent disposal solution: spent fuel is kept in pools and then dry casks at more than 70 sites in 35 states pending a federal decision.

“it could fit on a single football field at a depth of less than 10 yards”

“The newest US nuclear reactors, at Plant Vogtle in Georgia, came in years late and more than double their original budget.”

Confirmed: Vogtle Units 3 and 4 were expected to cost $14 billion and start in 2016 and 2017, but entered commercial operation in July 2023 and 2024 at an estimated total cost of more than $30 billion.

Construction began in 2009. Two other AP1000 reactors in South Carolina were abandoned in 2017, and when Vogtle 4 opened in 2024 no other reactors were under construction in the US. The cost figure is Georgia Power's estimate as reported by the EIA.

“Georgia Power now estimates the total cost of the project to be more than $30 billion”

“In the UK, Sizewell C will cost around £38 billion and will not produce power until the mid-to-late 2030s.”

Confirmed: the government's February 2026 summary business case puts Sizewell C's cost at around £38 billion in 2024 prices, with operations expected to begin in the mid-to-late 2030s.

The final investment decision was taken on 22 July 2025. Sizewell C, of up to 3.2 GW, is an above-ground replica of Hinkley Point C financed through a regulated asset base model, which the government says adds an average of about £1 a month to household bills during construction. In May 2026 the National Audit Office reported, in 2024-25 prices, a company baseline estimate of £38.2 billion, a government threshold of £40.5 billion close to the middle of the company's forecasts, and funding committed up to £47.7 billion, with first power expected from 2038.

“It is expected that the overall cost of the project will be around £38bn (2024 prices), and operations will begin in the mid-to-late 2030s.”

“The cost of solar power has collapsed, falling by nearly 90% in about 15 years.”

Confirmed: electricity from utility-scale solar PV cost $496 per MWh in 2009 and fell by 88% within 15 years.

These are global average generation costs without subsidies; they exclude the extra storage, backup and grid costs of managing variable output. Over the same period the cost of electricity from new nuclear rose.

“Electricity from utility-scale solar photovoltaics cost $496 per MWh in 2009. Within 15 years the price declined by 88%”

“Renewables now generate more than half of the UK's electricity.”

Confirmed: renewables supplied a record 52.5% of UK electricity generation in 2025, the second year in a row in which they produced more than half.

Gas was still the largest single source, at 31.5% of generation in 2025, and nuclear output fell 12% to 35.9 TWh because of outages across the ageing fleet. Shares vary with the weather and new capacity: provisional figures put renewables at 53.1% of generation in the first quarter of 2026, helped by record wind generation.

“Renewable’s share of electricity generation was a record 52.5 per cent in 2025”

“Hinkley Point C was guaranteed a price of £92.50 per megawatt-hour for its electricity.”

The 2016 contract set a strike price of £92.50/MWh, but in 2012 prices, and it was cut to £89.50 once Sizewell C reached its final investment decision in July 2025. The National Audit Office puts that at about £129/MWh in 2025 prices.

The strike price is fixed in real terms for the first 35 years of operation, so quoting £92.50 next to figures in current prices understates it. EDF's 2025 results confirm the £3/MWh reduction to £89.50/MWh.

“DESNZ guaranteed that HPC would receive £89.50 per megawatt-hour (MWh) in 2012 prices (£128.94 in 2025 prices) for the first 35 years”

“The Chernobyl disaster killed hundreds of thousands of people.”

Fewer than 50 deaths had been directly attributed to Chernobyl radiation by mid-2005. The WHO estimated up to 4,000 eventual deaths among the most exposed groups and up to about 9,000 including people in less contaminated areas.

These long-term figures are predictions based on a model that assumes even low doses carry some cancer risk. The WHO's cancer research agency estimated about 16,000 cancer deaths across Europe by 2065 (uncertainty range 6,700 to 38,000); mainstream scientific estimates do not approach hundreds of thousands. By 2006 nearly 5,000 thyroid cancers had been diagnosed in people who were children or adolescents at the time; in 2005 the WHO said survival among such patients had been almost 99%, judging from Belarus.

“A total of up to 4000 people could eventually die of radiation exposure from the Chernobyl nuclear power plant (NPP) accident”

“Radiation from the Fukushima accident killed thousands of people.”

One death, a worker's lung cancer, has been officially attributed to radiation from Fukushima. About 2,313 deaths were linked to the stress and disruption of the evacuation rather than to radiation exposure.

The earthquake and tsunami that caused the accident killed far more people than the nuclear accident. WHO and UN scientific assessments found very low radiation risk to the public; the evacuation itself caused most of the accident-related deaths.

“In 2018, the Japanese government reported that one worker has since died from lung cancer as a result of radiation exposure from the event.”

“Small modular reactors are ready now and can be rolled out in the UK within a few years.”

The UK's first SMR project, announced in November 2025 for Wylfa in North Wales with an initial three units, only aims to supply power to the grid from the mid-2030s.

In April 2026 Great British Energy - Nuclear signed a contract with Rolls-Royce SMR to begin site-specific design, regulatory and planning work ahead of a future final investment decision, backed by £2.6 billion allocated in the 2025 Spending Review. Factory-built SMRs are meant to shorten construction, but the Wylfa units would be the UK's first.

“small modular reactors to be supplying power to the grid from the mid-2030s”

“New UK offshore wind at £91.20/MWh costs about the same as Hinkley Point C at £92.50/MWh.”

The two figures use different price bases. In Allocation Round 7 (results published January 2026), offshore wind in England and Wales cleared at £65.45/MWh in 2012 prices (£91.20 in 2024 prices). Hinkley's £92.50/MWh is in 2012 prices and has since been cut to £89.50, so like for like new offshore wind is about 27% cheaper per MWh (our calculation from the government and National Audit Office figures in 2012 prices).

In today's money the National Audit Office puts Hinkley's price at about £129/MWh (2025 prices). Strike prices leave out system costs: the NAO notes that intermittent renewables need extra transmission, reserve generation and balancing services that standard cost measures do not capture, so a per-MWh comparison flatters wind somewhat. Contract lengths also differ: 20 years for the new wind contracts against 35 years for Hinkley.

“Breakdown of the outcome by technology, year and clearing price (strike prices are in 2012 and 2024 prices)”

“Nuclear is the cheapest way to generate electricity.”

Not for new plants: global cost data show electricity from new nuclear has become more expensive since 2009, while new solar and wind have become much cheaper and in most places now undercut new fossil fuels.

Rising nuclear costs are linked to tighter safety regulation and small supply chains after decades of few builds; countries that standardised designs kept costs steadier. Paid-off existing plants can run cheaply, and simple cost comparisons miss nuclear's value as steady output.

“For nuclear power, you see the data since 2009 in the chart. Nuclear power has increased in price.”

“In the UK, net migration hit a record high of around 944,000 in the year to March 2023.”

Confirmed: official estimates show UK net migration peaked at 944,000 in the year ending March 2023, the highest level on record.

Net migration is immigration minus emigration. The post-Brexit spike was driven by rising immigration of non-EU citizens; net migration has since fallen sharply, to 171,000 in calendar year 2025.

“when net migration peaked at a historical high of 944,000”

“In the UK, net migration is now the only source of population growth because deaths are about to outnumber births.”

Confirmed: ONS projects net migration adding 2.2 million people in the 10 years to mid-2034, while natural change (births minus deaths) is minus 450,000, with deaths exceeding births from mid-2026.

From the ONS 2024-based national population projections (April 2026), which assume long-term net migration of 230,000 a year. Projections rest on assumptions, and migration is the most uncertain component.

“It is projected that net migration will add 2.2 million to the population in the 10 years to mid-2034.”

“Most people in Britain think immigration has been too high.”

Confirmed: in survey data cited by the Migration Observatory, 69% said immigration over the past 10 years had been too high, although only 41% said it had been bad for Britain.

Answers depend heavily on question wording: many more people say numbers are too high than say immigration is harmful, and views of small-boat arrivals are far more negative than of visa routes. Immigration salience reached 51% naming it an important issue in September 2025.

“69% said it had been too high. This suggests that many people would prefer lower levels of immigration”

“Immigration pushes up house prices in the UK.”

There is some evidence for this: the Migration Advisory Committee (2018) found a 1 percentage point rise in the UK population from migration was associated with roughly a 1% rise in house prices, but warned the result depends heavily on the statistical method.

Migration adds to housing demand through population growth while supply has lagged, but prices are also driven by incomes, interest rates and credit. One local study (2003 to 2010) found migrant inflows lowered local prices as higher-income residents moved out, so the evidence is not one-directional.

“there is some evidence that migration is likely to have increased house prices in the UK”

“Some migrants, such as low-paid care workers, cost the UK public finances more over their lifetimes than they pay in.”

Confirmed: the Migration Advisory Committee (2025) estimated that care workers, Health and Care visa dependants and partners of British citizens have a net negative lifetime fiscal impact, and the OBR found low-wage migrants have a negative lifetime impact.

Fiscal impact depends mainly on earnings. The same MAC work found Skilled Worker and Health and Care main applicants make a positive lifetime contribution, and the OBR found an average-wage migrant arriving at 25 contributes more than a UK-born worker on the same pay because the UK did not fund their childhood education.

“care workers, Health and Care dependants, and the partners of British citizens have a net negative lifetime fiscal impact”

“Tens of thousands of people still cross the Channel to the UK in small boats every year.”

Confirmed: 33,000 people arrived by small boat in the year ending June 2026, 88% of the 38,000 detected arrivals via illegal entry routes.

Small-boat arrivals were 23% lower than the previous year but similar to the years ending June 2022 and June 2024. In the same year 86,000 people claimed asylum (down 21%), and 16,000 supported asylum seekers were in hotels at the end of June 2026, half the number a year earlier.

“there were 38,000 detected arrivals via illegal entry routes in the YE June 2026 - small boat arrivals accounted for 33,000 (88%) of these”

“The OBR says higher net migration reduces UK government borrowing.”

In its March 2024 forecast the OBR estimated that about 350,000 extra net migrants would cut borrowing by around £7.4 billion by 2028-29, mainly through £6.2 billion of extra tax. This is a five-year effect that weakens as migrants age and gain access to benefits.

The OBR also noted that, with departmental budgets fixed, extra migration lowers public service spending per person (by about £40 in 2028-29) and adds pressure on services. Long-run effects depend on migrants' earnings and length of stay.

“deliver a net reduction in borrowing of around £7.4 billion by 2028-29, taking account of lower spending on debt interest as well”

“In the US, undocumented immigrants are arrested for violent crimes at lower rates than native-born citizens.”

Confirmed: a study of Texas arrest data from 2012 to 2018 found undocumented immigrants were arrested at less than half the rate of US-born citizens for violent and drug crimes.

This is US evidence (mainly Texas) and does not directly describe the UK. PolitiFact also cites a National Bureau of Economic Research study covering 1870 to 2020 that found immigrants 60% less likely to be incarcerated.

“undocumented immigrants are arrested at less than half the rate of native-born U.S. citizens for violent and drug crimes”

“The NHS depends on overseas-trained staff, with around 40% of new nurses trained abroad.”

The figure is out of date: 39.1% of new joiners to the UK nursing and midwifery register were educated abroad in 2024-25, but only 26.4% in 2025-26.

These Nursing and Midwifery Council figures count nurses, midwives and nursing associates joining the UK register, not NHS staff specifically. The international share of new joiners peaked at 49.4% in 2023-24 and has fallen since. Reliance is still substantial: the NMC's full UK report shows 24.4% of the register was internationally educated in March 2026.

“In 2025-26, 26.4% of new joiners to the Register were internationally educated, compared to 39.1% the year before and a high of 49.4% in 2023-24.”

“The UK population is ageing, with more pensioners for every person of working age.”

Confirmed: ONS projects the old-age dependency ratio (people of pensionable age per 1,000 of working age) rising from 280 in mid-2024 to 310 in mid-2034 and 329 by mid-2049.

The pensionable-age population is projected to grow by 1.8 million to 14.2 million by mid-2034, even allowing for the State Pension age rising to 67. Most migrants arrive of working age, which is why debaters link migration to ageing, though migrants also age over time.

“It is projected that OADR will increase from 280 in mid-2024 to 310 in mid-2034, and reach 329 by mid-2049.”

“UK net migration has already roughly halved, to about 171,000 in 2025.”

Confirmed: provisional ONS estimates put net migration at 171,000 in the year ending December 2025, down from an updated 331,000 in 2024.

The fall was driven mainly by a 47% drop in non-EU nationals arriving for work. Immigration was 813,000 and emigration 642,000. These are provisional estimates published May 2026 and are routinely revised.

“At 171,000, long-term international net migration for year ending (YE) December 2025 has nearly halved from YE December 2024 (updated to 331,000)”

“Immigration has driven down wages for British workers.”

Research finds the effects on UK-born workers' wages are small: slightly negative for the lowest paid and slightly positive for higher earners. One study estimated immigration from 1994 to 2016 reduced hourly pay at the 5th percentile by about half a penny a year.

The Migration Advisory Committee's 2018 review found little or no impact on average wages or employment, with any negative effects concentrated among lower-educated workers and possibly during downturns.

“Research suggests that the impacts of migration on wages and employment prospects for UK-born workers are small”

“UK net migration topped one million a year.”

Net migration peaked at 944,000 in the year ending March 2023. The figure above one million was gross immigration, which peaked at about 1.47 million before emigration is subtracted.

Debaters often confuse immigration (people arriving) with net migration (arrivals minus departures). By 2025 net migration had fallen to 171,000.

“This was down from the peak of 944,000 in YE March 2023.”

“Most people who move to the UK are asylum seekers.”

Asylum seekers are a minority of arrivals. Of non-EU nationals immigrating long term in the year ending December 2025, 47% came to study, 23% for work and 14% were asylum applicants.

These are ONS provisional estimates for long-term non-EU+ immigration (627,000 people); family and humanitarian routes made up a further 7% and 6%. Total long-term immigration of all nationalities was 813,000, so the 88,000 non-EU+ nationals who came for asylum were about 11% of all arrivals.

“47% came for study-related reasons, 23% for work-related reasons, 14% were asylum applicants”

“Migrants can claim benefits as soon as they arrive in the UK.”

Most non-EU migrants on work, study or family visas are subject to the No Recourse to Public Funds condition, which bars most benefits such as Universal Credit and Child Benefit until they are granted indefinite leave to remain.

The rule does not cover everyone: for example, people granted refugee status or settlement can access benefits. The government says the condition is designed to prevent fiscal costs from benefit payments.

“prevents non-EU citizens on work, study, or family visas from accessing most benefits, including Universal Credit or Child Benefit”

“Migrants get priority for social housing over British people.”

Social housing is not allocated on the basis of immigration status or nationality, and at the 2021 Census non-UK passport holders were 7% of social housing residents in England and Wales against 10% of the population.

Migrants on temporary work, study or family visas are generally not eligible for social housing. Counting all foreign-born residents, including naturalised citizens, the share in social housing was 15% in 2021, and new lettings to non-UK nationals in England rose from 6% to 10% between 2007/08 and 2022/23.

“on average, they were slightly underrepresented among social renters at the time of the 2021 Census of England and Wales”

“The UK government has decided to ban social media for under-16s, with the ban expected to take effect in 2027.”

Confirmed: on 15 June 2026 the UK government announced that platforms will be blocked from offering services to under-16s. Regulations are due before Parliament by Christmas 2026 and are expected to take effect in spring 2027.

The UK says it will follow Australia's model, with Ofcom enforcing it. Messaging services such as WhatsApp and Signal are not intended to be covered. The ban was not yet in force as of September 2026.

“Social media platforms to be blocked from offering services to under-16s, marking a line in the sand and setting a new normal for future generations”

“Nine in ten UK parents support banning social media for under-16s.”

The nine in ten figure comes from the UK government's consultation, which drew more than 116,000 responses from people who chose to take part. In the government's own nationally representative survey, about three quarters of parents (76%) agreed social media should have a minimum age of at least 16.

The government's research report found 91% support among parents in the self-selecting public consultation, against 76% in its representative panel of 5,010 parents. A separate YouGov poll published on 15 June 2026 found 77% of parents of under-18s would support a ban and 14% were opposed. Support is high either way, but nine in ten overstates it.

“around three quarters of parents (76%) agree that social media should have a minimum age of access of at least 16”

“Teenagers who spend more than three hours a day on social media face double the risk of mental health problems.”

A US cohort study found that more than 3 hours a day was linked to about 1.6 to 1.8 times the risk of internalising problems (such as anxiety and depression) alone, and about 2 to 2.4 times the risk of combined internalising and externalising problems.

The study followed 6,595 US adolescents aged 12 to 15 (JAMA Psychiatry, 2019). 'Double the risk' fits the combined-problems result better than the anxiety and depression result alone. It is observational, so it shows an association, not proof that social media caused the problems.

“Adolescents who spend more than 3 hours per day using social media may be at heightened risk for mental health problems, particularly internalizing problems”

“When Facebook was rolled out across US colleges, students' mental health got worse.”

Confirmed: a natural experiment using Facebook's staggered arrival at US colleges found it had a negative impact on student mental health and increased reports of mental-health-related harm to academic performance.

Published in the American Economic Review in 2022, this is one of the strongest pieces of causal evidence for harm. It covers college students in the mid-2000s, not under-16s on today's platforms. The authors point to unfavourable social comparison as the likely mechanism.

“We find that the rollout of Facebook at a college had a negative impact on student mental health.”

“Problematic social media use among adolescents in Europe rose from 7% in 2018 to 11% in 2022.”

Confirmed: WHO/Europe's HBSC survey found problematic social media use rose from 7% in 2018 to 11% in 2022, with girls higher than boys (13% vs 9%).

The survey covered almost 280,000 young people aged 11, 13 and 15 in 44 countries and regions across Europe, central Asia and Canada. 'Problematic use' means addiction-like symptoms such as being unable to control use.

“rates increasing from 7% in 2018 to 11% in 2022”

“Australia's under-16 social media ban led to about 4.7 million accounts being removed within weeks of taking effect.”

Confirmed: the ban took effect on 10 December 2025, and Australia's eSafety Commissioner reported that about 4.7 million age-restricted accounts had been removed or restricted by mid-December 2025.

eSafety says the figure counts accounts, not children, because individuals often have several accounts. In March 2026 eSafety said it was investigating Facebook, Instagram, Snapchat, TikTok and YouTube for possible non-compliance, and on 31 July 2026 it said those investigations were still ongoing. An independent review of the law must be started within two years of it taking effect.

“about 4.7 million age-restricted accounts were removed or restricted from platforms as at mid-December 2025”

“In Australia, many under-16s still have social media accounts and some are getting around the ban.”

Confirmed: in July 2026 Australia's eSafety Commissioner reported that most under-16s who had social media accounts before the law took effect had kept them or created new ones three months later, with platforms' failure to use effective age checks cited as the main reason.

In eSafety's study of more than 4,000 children and families, the share of under-16s with a social media account fell from 52.4% to 42.1%, which eSafety called a modest but statistically significant decline. Its March 2026 compliance report found some platforms let under-16s repeat the same age check until they got a 16+ result. The UK says its own ban will require stronger age checks.

“most under-16s who had social media accounts before commencement were able to either retain them or create new ones at the three-month mark”

“In the UK, 81% of 10 to 12 year olds use social media even though most platforms set a minimum age of 13.”

Confirmed: Ofcom's 2025 research, cited by the UK government, found 81% of 10 to 12 year olds use at least one social media app or site, and 86% have their own account on a social media, messaging or video-sharing platform.

The figures come from Ofcom's survey of October and November 2024, in which children aged 8 to 17 answered for themselves. Most social media services that set a minimum age use 13, in line with the age of digital consent.

“found that 81% of 10–12-year-olds use at least one social media app or site”

“Far fewer teens say social media harms them personally than say it harms their peers, and most say it helps them feel connected to friends.”

Confirmed: in Pew's US survey, 14% of teens said social media has a negative effect on them, compared with 48% who said so about people their age. About three-quarters said it makes them feel more connected to friends.

The survey covered 1,391 US teens aged 13 to 17. The share saying social media is mostly negative for their peers rose from 32% in 2022 to 48%, so both sides can cite it.

“Far fewer teens say social media have a negative effect on them than say this of their peers (14% vs. 48%)”

“Across OECD countries, 15-year-olds who use social media moderately do better in maths than both non-users and heavy users.”

Confirmed: OECD analysis of PISA 2022 data found maths scores were highest for moderate users, lower for both non-users and heavy users, and fell steadily beyond three hours a day.

This is a correlation, not proof of cause and effect. The same OECD paper says the average 15-year-old spends almost 35 hours a week on social media, and that research on wellbeing is 'complex and nuanced'.

“Mathematics performance is highest among students who use social media moderately, with both non-users and heavy users performing worse.”

“Face-scanning age checks are least reliable for teenagers close to 16, the exact age a ban needs to police.”

Confirmed: Australia's government-funded age assurance trial found facial age estimation was less accurate for younger users and least certain near the 16+ and 18+ thresholds, as summarised by the UK government.

The trial's headline finding was that age assurance can be done privately and effectively, but that no single method suits every case. Checking a child's ID is also hard because many children have no official photo ID.

“have higher levels of uncertainty when determining the age of users close to the 16+ or 18+ age threshold”

“Science has proven that social media causes mental illness in children.”

The UK government's own evidence summary says the evidence on screen time and children's mental health is mixed and no causal link has been found. A 2025 review by UK academics led by Professor Amy Orben found associations but a lack of high-quality causal evidence.

No proven causal link does not mean there is no harm. Some quasi-experimental studies suggest harm (see SM-04), and the government says the absence of a causal finding does not rule one out.

“The evidence around screentime and the impact on children's mental health is mixed, and no causal link has been found.”

“Under Australia's ban, children or their parents are fined if kids use social media.”

Australia's law puts the burden entirely on platforms. There are no penalties for young people who get access, or for parents and educators who help them.

Platforms must take 'reasonable steps' to stop under-16s creating or keeping accounts, and the eSafety Commissioner enforces this. The UK has said the same for its planned ban: responsibility sits with platforms, not children.

“No penalties are foreseen for young people gaining access to social media, or parents and educators who may provide such access.”

“Australia's law completely blocks under-16s from seeing anything on platforms like YouTube or TikTok.”

Australia's law stops under-16s from having accounts on age-restricted platforms, not from viewing content. eSafety says under-16s can still see publicly available content that does not require logging in, such as most content on YouTube.

Platforms that fail to take reasonable steps to keep under-16s off can face court-imposed fines of up to 150,000 penalty units, which eSafety now puts at A$54.6 million. It was quoted as A$49.5 million before the Commonwealth penalty unit rose from A$330 to A$364 on 1 July 2026. Services whose sole or primary purpose is messaging or online gaming are excluded.

“Under-16s are still allowed to see publicly available social media content that doesn’t require being logged into an account.”

“UK VPN use doubled because children were using VPNs to get round online age checks.”

UK VPN use did more than double after mandatory age checks began on 25 July 2025, from about 650,000 daily users to a peak of over 1.4 million in mid-August 2025. But the UK government and Ofcom say the evidence does not show the rise was driven by children trying to get round age checks.

It is impossible to know how many VPN users are children. Ofcom's July 2026 report, using mobile app data from a different supplier, estimated 2.2 million daily UK VPN users against 1.2 million before the checks. Its March 2026 survey found a quarter of 11 to 17 year olds in Great Britain had used a VPN in the past six months, but only 5% had used one to reach content meant for older people or age-restricted features, which Ofcom says may be an underestimate. In July 2026 the UK government said it will not ban or age-gate VPNs.

“early evidence does not suggest the peak in VPN use was driven by children seeking to bypass age assurance processes.”

“Teen mental health in the US keeps getting worse year after year.”

CDC data show persistent sadness or hopelessness among US high school students rose from 30% in 2015 to a peak of 42% in 2021, then fell to 40% in 2023 and 33% in 2025. It is still worse than a decade ago but has improved recently.

The figures come from the CDC's Youth Risk Behavior Survey. The CDC describes the trend as improved over two years but worse over ten. Recent improvement weakens claims of a relentless decline, but not the concern about the longer-term rise.

“In 2025, 33% of high school students experienced persistent feelings of sadness or hopelessness during the past 12 months.”

“Alaska has paid every resident an unconditional cash dividend for decades, and it has not reduced employment.”

Confirmed: a study of the Alaska Permanent Fund Dividend (paid since 1982) found no effect on employment and a 1.8 percentage point (17%) rise in part-time work.

The dividend is modest, about $2,000 per person a year in recent years, and the authors note it may be smaller than a full basic income. They link the result to the cash boosting local demand. A 2023 peer-reviewed study found small cuts in women's hours and more employment among men after payouts, a net 0.8% annual labour market contraction per extra $1,000.

“we show that the dividend had no effect on employment, and increased part-time work by 1.8 percentage points (17%)”

“In Finland's basic income experiment, unemployed people given an unconditional monthly payment did not work less than those on normal benefits.”

Confirmed: 2,000 unemployed people were randomly given a basic income in place of minimum unemployment benefits, and their days in employment were statistically unchanged in the first year.

Finland ran the trial in 2017-2018, paying €560 a month with no job-search requirements. It cut participation tax rates for full-time work by 23 percentage points, yet first-year employment did not rise significantly. A 2020 evaluation by Finland's VATT institute found 6 more days of employment from November 2017 to October 2018, but an activation model introduced in early 2018 complicates that result.

“Despite the considerable increase in work incentives, days in employment remained statistically unchanged in the first year of the experiment.”

“In the UK, Universal Credit is cut by 55p for every extra £1 a claimant earns, which weakens the reward for working more.”

Confirmed: above any work allowance, Universal Credit falls by 55p for each £1 of earnings.

The taper was cut from 63% to 55% in November 2021. Claimants who qualify for a work allowance can earn £427 or £710 a month before the taper starts. Income tax and National Insurance come on top, so the effective loss on extra earnings is higher. Supporters of UBI argue that a flat payment would not be withdrawn as earnings rise.

“For every £1 you earn from working, your Universal Credit payment goes down by 55p.”

“Means-tested benefits miss many people who are entitled but never claim, while a universal payment would reach almost everyone.”

OECD modelling assumes a universal basic income would have near-complete take-up, unlike means-tested benefits. In the UK simulation, the poorest tenth of households gained, often because they were entitled to benefits they did not claim.

Near-universal take-up is a modelling assumption, not an observed UK result. The same OECD analysis found that a budget-neutral UK basic income would still leave many other low-income households worse off.

“Those in the lowest income decile would also gain, as these are often households that are entitled to means-tested benefits but do not claim.”

“The OECD estimates that automation could wipe out about 14% of jobs and radically change another 32%.”

This was the OECD's 2019 estimate: 14% of existing jobs could disappear through automation in the next 15-20 years and another 32% were likely to change radically. A 2023 OECD update that includes AI puts about 27% of employment in occupations at highest risk.

The 2019 figures are OECD averages; the share of high-risk jobs ranged from 6% in Norway to 34% in the Slovak Republic. OECD researchers say the higher recent figure may reflect AI progress and a different method, and does not mean the risk of automation has doubled. In 2019 the OECD said mass technological unemployment seemed unlikely.

“14% of existing jobs could disappear as a result of automation in the next 15-20 years, and another 32% are likely to change radically”

“A UK basic income paid at the level of existing minimum benefits would cost far more than the benefits it replaced.”

Confirmed: OECD modelling found that such a scheme left a £44.3bn a year gap, even after scrapping most working-age benefits and the tax-free personal allowance. Closing the gap would take income tax rates about 25% higher.

This is a 2017 simulation by the OECD Secretariat using figures for 2015. The 25% is a proportional rise in all income tax rates (for example 20% to 25%), not 25 percentage points. In Finland and France, by contrast, the same design was close to budget-neutral.

“In the UK, however, a BI would be much more expensive than the existing benefits it would replace”

“A UBI the UK could afford without raising taxes would pay adults only about £230 a month.”

OECD modelling found a budget-neutral UK basic income, paid for by abolishing most working-age benefits and the tax-free personal allowance, would be £230 a month per adult and £189 per child after tax, in 2015 terms. The OECD says this is 28% below the level of existing minimum-income benefits.

The OECD itself describes abolishing tax-free allowances as a tax increase, so this is not a scheme without tax rises. A basic income at minimum-benefit level would pay £317 a month per adult after tax. Under the £230 version, lower-income UK households, including many lone parents, often lose out, though the very poorest non-claimants gain.

“A revenue-neutral scheme would require the BI amount to be reduced to 28% below GMI levels.”

“The IMF estimates that a basic income worth a quarter of median income would cost about 6.5% of GDP in rich countries.”

Confirmed: the IMF's October 2017 Fiscal Monitor put the gross cost of a UBI at 25% of median per capita income at about 6.5% of GDP in the average advanced economy and 3.75% in emerging markets.

This is a gross cost on top of existing programmes. It ignores how the UBI would be financed and any behavioural responses. A payment at 25% of median income is far below a poverty-line income.

“A UBI set at 25 percent of median per capita net market income would cost about 6½ percent of GDP and 3¾ percent of GDP”

“The IMF has concluded that in rich countries with generous welfare systems, a UBI is unlikely to work well as a replacement for existing benefits.”

Confirmed: the IMF said a UBI is unlikely to be an effective substitute where safety nets are generous and progressive. It advised fixing coverage gaps and take-up first.

The IMF saw a stronger case for a UBI in emerging and developing economies whose safety nets have large coverage gaps and low progressivity, if it can be financed efficiently. It also noted that digital ID and payment technology make a UBI easier to deliver.

“In advanced economies, where existing safety nets are often generous and progressive, a UBI is unlikely to be an effective substitute.”

“A basic income funded by scrapping tax-free allowances would sharply weaken work incentives for people whose partner already works.”

Confirmed: OECD modelling found this design would severely weaken work incentives for second earners in almost all countries studied. It would strengthen them for many lone parents and single-earner couples who now face benefit withdrawal.

Second earners lose their tax-free allowance and gain little in return, because they were not receiving means-tested benefits anyway. The overall effect on work depends on the design and how it is financed.

“A BI as outline above would severely weaken work incentives for those whose partner is in paid work.”

“Giving every UK adult a basic income of around £9,600 a year would cost roughly £500 billion a year.”

Confirmed as arithmetic: Full Fact calculated that £504bn is equivalent to about £9,600 a year for each UK adult aged 18 or over, before administration costs.

This is a 2019 gross figure, before savings from replaced benefits or extra tax. The ILO then put the UK poverty threshold at about £8,800 a year per adult. The £504bn came from unpublished Conservative analysis of Labour plans, and Full Fact noted that Labour had only pledged trials.

“£504 billion is the equivalent of providing a universal basic income of around £9,600 a year to each adult aged 18 or over”

“A universal basic income would end poverty in the UK.”

OECD modelling found that a budget-neutral UK basic income would raise working-age poverty. Even a version paid for by much higher income taxes did not significantly reduce it.

Flat payments spread money across everyone, including those who are not poor, and replacing targeted benefits leaves some people worse off. A UBI set above the poverty line could cut income poverty, but at a far higher cost. The result depends on the payment level and how it is funded.

“it would not prove to be an effective tool for reducing poverty”

“Automation is set to wipe out around half of all jobs.”

The OECD's 2023 estimate, which accounts for AI, puts about 27% of employment in occupations at highest risk of automation; its 2019 estimate put 14% of jobs at high risk. The 'almost half' figure (47%) is Frey and Osborne's estimate for US jobs.

Being at high risk does not mean a job will disappear. In 2023 the OECD found little evidence of significant job losses from AI so far, though most of that research predated the latest generative AI. ONS analysis put 7.4% of jobs in England at high risk of automation in 2017.

“Taking the effect of AI into account, occupations at highest risk of automation account for about 27% of employment.”

“Finland has introduced a universal basic income for all its citizens.”

Finland ran a two-year trial from 2017 to 2018. 2,000 unemployed benefit recipients received €560 a month in place of their minimum unemployment benefit. It was never a national, universal scheme.

The trial covered only unemployed people chosen at random, so it cannot show the cost, tax or work effects of a truly universal payment.

“2,000 unemployment benefit recipients were randomized to receive a monthly basic income”

“People given unconditional cash mostly waste it on alcohol and tobacco.”

A review of 19 quantitative studies by World Bank researchers found that, almost without exception, cash transfers had no significant effect on spending on alcohol and tobacco or reduced it. The only two studies that found a significant rise were non-experimental, and the effects were small.

The evidence comes from cash transfer programmes in developing countries in Latin America, Africa and Asia, not from rich-country basic income trials. The authors found the result held for both conditional and unconditional programmes.

“Almost without exception, studies find either no significant impact or a significant negative impact of transfers on temptation goods.”

“Basic income trials show that giving people free money does not make them work any less.”

Some schemes found no fall in employment, including Alaska's dividend and Finland's trial. A large US randomised trial paying $1,000 a month for three years found income excluding the transfers fell by about $1,900 a year, labour market participation fell 4.2 percentage points and work fell by 1-2 hours a week.

The US study randomised 1,000 low-income people to get $1,000 a month and 2,000 controls to get $50 a month, in Illinois and Texas. Partners also cut their hours, and leisure rose most. Because it targeted low-income people, it is not a test of a universal payment; the Alaska study links its result to the cash boosting local demand.

“The transfer caused total individual income excluding the transfers to fall by about $1,900/year relative to the control group”

Homepage demo · checked 24 Sept 2026

Legal sources took about 72% of cannabis spending in Canada in 2025, up from 22% in late 2018.

Homepage demo · checked 24 Sept 2026

A 47-state study of nearly 900,000 US high-schoolers found no net rise in teen cannabis use after legalisation. Other studies found small rises, so the evidence is mixed.

Wider evidence: fact card CB-14

Homepage, “The judge” section · checked 24 Sept 2026

UK net migration peaked at 944,000 in the year to March 2023. No official estimate has topped a million.

Motion: nuclear power, Pro side summary · checked 24 Sept 2026

Nuclear is one of the safest energy sources per unit of electricity, with death rates similar to wind and solar.

Wider evidence: fact card NU-01

Motion: free trade, Pro side summary · checked 24 Sept 2026

Between 1990 and 2019 the number of people in extreme poverty fell by about 1.3 billion.